Our first human therapeutic trial with a DNA plasmid construct containing HIV-1 genes env and rev showed safety and evidence of both cellular and humoral immune responses, but no consistent impact on HIV surrogate markers. Our second trial is a randomized double-blind vehicle-controlled does-ranging study giving 2 constructs containing env/rev and gag/pol genes to HIV-positive subjects whose viral production is suppressed by HAART. Subjects with CD4 cell counts >400 and RNA copy number (viral load :VL:)) >1,000/mi are given 3 intramuscular injections at 4 week intervals with a booster at 24 weeks; subjects are randomized 5:2 to receive vaccine: diluent. As of 10/1/98 the 100ug cohort has completed its primary series and booster, and the 300ug cohort is fully enrolled (7 subjects). Mean entry CD4 count was 633 and VL <400 in 12, <1000 in all. Minor local and no systemic clinical toxicity has been noted. CBC, urinalysis, and serum chemistries remained normal except for 1 subject whose creatine kinase transiently rose to twice normal. None have developed antinuclear or antiDNA antibodies. VLs have remained <400 in 12 and became <400 in 1 of 2 initially detectable. PBMC ultrasensitive HIV DNA assay (Roche) has shown no clear trend to date. In the 100ug cohort, lymphocyte stimulation index to rev antigen increased by >=5 in 6/7 subjects, often at several time points: Sl to p24 increased by >5 in 5/7 subjects, and to gp 160 in 3/7. Initial data from the 300ug cohort shows responses in 3/5 by week 8. CTL assays and B-chemokines are pending. This on-going study demonstrates that simultaneous administration of two HIV DNA constructs appears safe and may induce immune responses in HIV-infected subjects receiving highly effective antiviral therapy.